►
From YouTube: DataHub @ LinkedIn: Extending the OSS UI
Description
Aikepaer Abuduweili & Joshua Shinavier from LinkedIn give an overview of how they have begun extending the DataHub OSS UI during the November 2021 Community Town Hall.
A
One
more
session
from
the
linkedin
community
giving
us
an
update
on
what
they've
been
doing
with
extending
the
data
hub,
ui,
so
abu,
I
will.
A
B
Share,
okay,
so
the
this
is
just
going
to
be
a
quick
intro
for
me
and
I'll
hand,
things
back
over
to
abu,
so
really
exciting.
Things
happening
on
the
open
source
side
of
things.
We're
going
to
tell
you
a
little
bit
about
some
exciting
developments
happening
on
the
linkedin
side.
Who
am
I?
My
name
is
josh
genevere.
If
you've
heard
of
me
or
you
recognized
my
zoom
background,
it
probably
had
something
to
do
with
knowledge
graphs.
B
If
you
want
to
know
more
about
why
I'm
excited
about
datahub
and
what
brought
me
here,
I'd
be
happy
to
chat
with
you
on
slack
but
yeah
yeah.
We
got
just
10
minutes,
I'm
going
to
hand
things
right
back
over
to
abu
and
then
I'll
be
back
at
the
end.
To
tell
you
a
little
bit
about
kind
of
vision
and
goals.
A
Hey
everyone,
my
name
is
abu,
I'm
a
ui
architect
from
linkedin.
So
so
you
guys
know
like
data
hub
is
an
internal
project
we
incubate
in
linkedin.
Then
we
open
source
that
to
a
community,
but
for
past
couple
years
we're
still
maintaining
our
internal
version
because,
like
we
still
have
internal
ask
or
requirement
so
so
we
so.
The
both
applications
share
a
lot
of
similar
features,
but
we
go,
but
it
goes
a
different
direction.
A
A
So,
under
that
contacts
we
is,
we
continue
evolving
our
platform
and
we
got
a
lot
of
ask
say:
hey:
can
we
extend
our
model
or
hey?
Can
I
add
a
new
tab
in
the
data
data
sets
page,
or
can
I
integrate
the
data
hub
to
our
internal
workflow
and
provide
the
seamless
experience?
So
we
are
thinking
like
what
is
the
next
generation
of
data.
A
Hadoop
will
be
so
under
the
contacts
we
are
thinking
is:
maybe
you
can
shift
data
hub
from
a
product
to
a
platform
and
enable
the
whole
extensibility
and
the
scalability,
so
once
I'm
introduced
to
a
new
data
hub
is
a
microphone
architecture,
so
micro,
front-end
architecture
is
a
design
principle,
design
philosophy
to
bring
multiple
teams
and
application
together
and
provide
one
seamless
ui
to
our
end
user.
So
you
can
see
like
at
the
right
size.
A
It
shows
the
evil
evaluation
of
a
microfront
so
like
a
frontal
analytics
to
macro
services
to
to
the
micro
front
end
and
the
micro
services.
So
this
enables
the
ui
the
first
one
is
enable
ui
extension,
so
extension
can
happen
in
the
application
level.
What
can
happen
in
the
page
level,
even
in
the
component
level?
So
shortly
I
will
demonstrate
this
part,
and
second,
one
is
that
it
can
provide
a
seamless
integration
to
other
application,
so
we're
going
to
leverage
the
deep
link
and
also
the
component
reuse,
to
make
the
experience
better.
A
Third
one
I
believe
most
of
us
has
been
through
the
application,
refactoring
or
application
upgrade,
and
the
you
guys
know
like
before.
You
have
a
full
feature
party.
You
are
not
able
to
release
that
application,
but
with
the
micro
front
end,
actually
you
can
release
your
application
page
by
page
or
feature
by
feature,
or
even
you
can
set
target
your
a
new
feature
to
a
certain
group
of
people
and
last,
not
least,
is
to
enable
the
cross
application
component
reuse.
It
is
really
tricky
in
the
ui
development
work
like
hey.
A
I
want
to
reuse
some
component
from
another
application.
It
is
really
tricky
and
especially
when
it's
become
the
the
functional
components
means
like
it's
connected
back
in
or
connected
to
apis,
it's
even
harder,
but
there
is
a
microphone,
tends
to
enable
the
whole
magic.
So
today,
I'm
going
to
use
like
one
demo
to
showcase
all
the
capability,
so
this
demo
is
to
showcase
how
we
migrate
from
the
amber
to
react
and
the
the
basic
pro
I
just
draw
like
a
very
simple
diagram.
A
So
the
basic
principle
is
like
the
yellow
box
is
the
new
application
we
are
building.
At
the
same
time,
we
also
host
our
existing,
amber
app
and
based
on
the
different
route
will
be
based
on
a
different
criteria.
We
route
the
traffic
to
the
data
to
the
new
application
and
we
can
turn
on
and
off
that
anytime.
You
elect
and
another
part
is
we.
We
don't
want
to
reinvent
the
view
and
there
are
a
lot
of
great
innovation
happening
in
the
open
source
site.
A
So
we
are
saying
like
hey:
can
we
reuse
any
things
from
the
open
source
site
and
you
and
use
that
internally?
So
this
also
will
demonstrate
how
this
can
happen,
and
that's
really
great.
So
I
will
switch
to
the
demo
so
here.
So
this
is
our
previous
data
application.
You
can
see
we
have
as
built
by
amber-
and
this
is
all
the
good
ui
and
the
one
thing
we
like
to
show
like.
A
So,
let's
see
we
want
to
switch
the
data
set
to
a
new
ui
so
like
from
the
previous
application,
I
go
to
the
data
set
page
it
go
to,
it
will
switch
to
the
new
page
and
the
two
thing
on
this
page
is
really
cool.
One
is,
if
you
look
at
this
section
it
does.
It
looks
familiar
so
this
is
a
component.
We
reused
from
the
open
source
site
and
the
cool
part.
A
Is
we
host
the
open
source
data
hub
internally
and
expose
the
component
directly
to
the
ui,
so
it's
happening
in
the
browser
side.
It
doesn't
go
through
the
whole
pipeline
or
like
at
the
whole
like
npm
route.
So
anytime
we
have
any
time
we
update
the
open
source.
It
will
automatically
upgrade
over
the
new
application.
Another
one
is
like,
as
I
mentioned
like
you,
can
simply
integrate
to
another
data
management
lifecycle.
So
like
oh,
I
want
to
go
from
the
catalog
or
to
our
reporting,
or
I
want
to
go
to
the
experiment.
A
I
can
directly
go
to
that
and
also
it
still
carry
all
the
contacts
to
the
new
application.
So
this
is
a
like
a
really
cool
part
under
and,
as
I
mentioned
like
a
component
review,
this
can
come
from
the
open
source
data
hub,
but
also
it
can
come
from
the
other
application.
So,
let's
see
if
some
certain
team
wants
to
extend
this
data
set
and
they
can
still
build
the
application
and
register
to
the
data
hub
and
expand
that,
so
that
is
a
really
cool
part.
A
Another
one
is
like
so
so
this
is
our
traditional
profile
page
and,
let's
see
tomorrow,
we
want
to
switch
that
to
a
new
page
literally.
You
just
need
to
do
one
line
of
change.
A
So,
let's
see
here,
I
I
changed
my
routing
and
come
to
this
page
then
need
to
go
to
the
new
application
and-
and
then,
let's
see
if
if
one
day
we
try
to
hey
the
full
application
is
ready,
we're
ready
to
switch
over
like
flip
the
switch
so
another
one
line
of
change,
then
boom
we
go
to
the
new
application.
B
I
mean
thanks
a
lot
about
you,
flip
to
the
next,
the
last
slide
yeah.
So
just
two
minutes
left
yeah
thanks
for
that
that
great
demo,
I
just
want
to
talk
a
little
bit
about
kind
of
the
the
big
picture.
Abu
mentioned
sort
of
the
split
between
open
source
data
hub
and
linkedin
data
hub.
You
know
on
the
open
source
side,
you
have
sort
of
the
community
engagement
that
is
really
attractive
to
us.
B
On
our
side
we
kind
of
set
the
bar
really
high
in
terms
of
scalability
and
in
terms
of
having
solutions
for
problems
that
you
know,
bigger
companies
have
things
to
do
with
compliance
and
so
on.
So
I
think
every
small
company
wants
to
grow
into
a
big
company
and
there's
there
is
some
value
into
in
sort
of
propagating
some
of
our
solutions
up
into
open
source.
So
you
know
we
will
continue
migration
to
react
and
graphql.
B
We
want
to
make
it
easier
for
us
to
contribute
code
into
open
source
data
hub
and
you
know,
consume
community
contributions.
There's
some
good
work
to
be
done.
I
think
around
interoperability.
Thinking
about
you
know
pdl
versus
graphql
as
a
modeling
language
supporting
grpc.
B
That's
what
I've
seen,
I
think,
there's
a
balance
to
be
struck
between
the
flexibilities,
which
is
what
we
see
on
the
open
source
side
and
then
robustness,
which
is
something
that's
very
important
to
us
on
the
linkedin
side,
having
strong
types
having
you
know,
strong,
schemas,
being
able
to
perform
inference
and
their
advantages
in
terms
of
optimizations
that
you
can
think
about
as
well,
and
I
just
added
metadata
knowledge
graphs,
because
that's
kind
of
my
my
bias,
I
see
a
lot
of
companies
building
large
scale,
knowledge
graphs
and
they
really
don't
have
a
good
solution
for
managing
their
metadata.
B
And
so
I
see
you
know,
data
hub
is
potentially
you
know,
leading
the
pack
as
a
data
catalog
for
enterprise
knowledge,
graphs,
and
that
is
about
it.
Unless
there
are
any
questions.